Q1.What level of accuracy can I expect?

Accuracy is highly dependent on the quality of the hard copy document you scan. The best OCR programs can reach even 100% accuracy when reading good quality printouts, typewritten texts, textfiles. An independent test conducted on multilingual texts declared Recognita Plus the most accurate OCR. Accuracy of Recognita is increased by further enhancements to the proven and successful two-step reading with Self Assertion Technology. (SAT), and proofing is aided by the use of dictionaries.

Q2.For what application would I consider using OCR?

There are numerous applications where fast, efficient text input is required, such as wordprocessing, desktop publishing, document image processing, document archiving, form processing, etc. It is a useful tool in translating written text input to voice output in reading systems.

Q3.What types of document can I read with an OCR? Can it recognize hand-written and foreign language text?

Good quality OCRs read complex page layouts with grayscale and/or line-art images produced by typing, typesetting, laser printing, matrix printing and their good quality copies. These include books, magazines, leaflets, business letters, financial, legal documents, statistical tables and so on. The recognition of handwritten text is usually not part of OCR programs, though Recognita Plus can recognize hand-printed numerals. Leading OCRs recognize foreign language letters like diacritical (accented) ones or Greek. Some of them even allow mixing of several languages on the same page.

Q4.What hardware do I need to operate OCR?

OCRs differ in hardware requirements. Some of them even run on a base-line personal computer configuration (like IBM PC/AT with 1 MB RAM and hard disk) using some image digitizing device (scanner). Others demand Megabytes of memory and/or more expensive processors (386/486/Pentium). There are also OCRs which utilize special boards with additional processor.

Q5.Will Recognita Plus work with Macintosh?

Although you cannot install Recognita Plus directly on your Macintosh, it can generate so many output file formats on a PC configuration that you must find one you can convert to Macintosh formats.

Q6.Is there any other software I have to buy to use my OCR?

Typically not, although a few OCR need some image processing software to get digitized files, while others require MS-Windows.

Q7.Will Recognita Plus work with several scanners?

Yes, Recognita, can take input from some 140 scanner models (including the TWAIN interface), or from image files in several formats, so you can be almost 100% sure that it will run together with yours, too. AccuPage 2.0 is supported, as is grayscale scanning with auto-threshold, auto-deskewing and image rotation to give maximum input image quality.

Q8.Are there any advantages for me in buying a 400 dpi scanner instead of a 300 dpi scanner?

Yes, if the scanner has a real image resolution of 400 dpi and not a software emulated one. Real 400 dpi scanners help to read small letters better.

Q9.Should I buy a hand-held or a full-page scanner?

For professional use (e.g. more than 5-10 pages a week) you definitely have to buy a full-page scanner. If you want to achieve nearly the same accuracy with a handheld scanner, your choice of font independent OCR is quite limited, check only the latest ones (e.g.: Recognita GO-CR).

Q10.Can I scan a full page with a hand-held scanner and Recognita Plus?

Basically, if the scanning software can create a full image having scanned the original in some steps, you can, but Recognita offers you a more adequate solution for hand-held scanners: Recognita GO-CR, and also a complete solution called Roll-And-Read Personal (including a hand-held scanner and Recognita GO-CR).

Q11.How can I find out whether the OCR I am considering for purchase meets my expectations?

Font independent software only OCR packages, especially those with added trainability do produce good result with most of the typical office documents. There are three or four leading packages of this kind. Narrow the choice by deciding on questions like "Can I really cope with my applications without the added benefits of trainability?" or "Can I drop recognition of foreign languages text from my list of requests?" or "Do I want to upgrade or replace my PC just for being able to use OCR?" or "Beside accuracy, how important is the speed of recognition for me?" or "Do I have a reasonable freedom of choice with supported scanners?" Soon you will be down to one or two OCR packages.

Q12.How do OCR programs work?

OCR programs process bitmaps taken directly from scanners or from image files. They isolate the textlines and determine the shapes to be recognized as characters (character separation phase). There are two basic kinds of algorithms used for the character recognition phase: matrix matching method tries to compare the character bitmaps to pre-stored patterns (it needs separate patterns to each style and size requiring pre-training and/or lot of memory) and feature analysis method which looks for special characteristics of the shapes like holes, edges, angles, etc. The latter approach results in font independent recognition. The page decomposition phase serves for providing format information of the original layout like paragraphs, columns, margins, indents, etc. For more information see OCR - What It Is And How It Works?

Q13.Does Recognita Plus handle images/pictures?

Yes, you can process complex page layouts with Recognita Plus. Images are retained in the full Recognita internal format. This is especially useful for handling lineart graphics and graphs, photos are converted to b/w format. From the full format you can even drag-and-drop the images to other applications, or save it in Recognita format together with text.

Q14.Can I process forms with Recognita Plus?

Recognita Corporation offers a special solution for that: Recognita Form. It reads data from forms by OCR, ICR, OMR and barcode technology. Forms are read into unlimited recognition zones and up to 255 datafields.

Q14.Can my OCR take over files from other image processing programs and can I expect it to give me image files that other programs can use?

Yes, most OCR can read different typical image files (like TIFF). Many of them also write such files that then can be taken as input to other imaging software.

Q15.Do OCR programs support my word processor, spreadsheet and other office applications efficiently enough?

The leading OCR software products offer 20 to 30 output file formats including known word-processor, spreadsheet and database ones. They usually give good formatting information together with the text, like indentation, alignment, column separation, table structure, etc. Check the list for the one you use.

Q16.What can I expect from my OCR if I want to integrate it in sophisticated applications, like optical filing, document processing and voice output systems?

Most OCR vendors consider this as a separate market and offer separate products for it. Few of them have multiplatform policy and support for more operating environments and also offer some kind of quasi application programmers' interface for integration of their OCR in optical filing, document processing, voice output reading systems.